On 02.10.2017 12:20, Tanu Kaskinen wrote: > On Sat, 2017-08-19 at 17:48 +0200, Georg Chini wrote: >> This patch extends the client subscription API, so that signals sent from >> PulseAudio can be processed. Within PulseAudio, a signal can be emitted >> using pa_signal_post(). The interface can be used to notify the client of >> events that are not covered by the subscription API (for example a button >> press event on a bluetooth headset). >> >> Setting up signal notification is very similar to using subscriptions. >> First the client needs to subscribe with pa_context_subscribe_signals() >> and then sets up a signal handler using pa_context_set_signal_callback(). >> >> The signal handler will receive three arguments in addition to the usual >> context and userdata: >> sender - string that specifies the origin of the signal >> signal - string that specifies the type of the signal >> signal_parameters - optional string for additional information >> --- >> PROTOCOL | 3 ++ >> src/map-file | 2 ++ >> src/pulse/def.h | 4 +++ >> src/pulse/internal.h | 2 ++ >> src/pulse/subscribe.c | 66 +++++++++++++++++++++++++++++++++++++---- >> src/pulse/subscribe.h | 9 ++++++ >> src/pulsecore/core-messages.c | 23 ++++++++++++++ >> src/pulsecore/core-messages.h | 11 +++++++ >> src/pulsecore/core.h | 1 + >> src/pulsecore/native-common.h | 1 + >> src/pulsecore/pdispatch.c | 1 + >> src/pulsecore/protocol-native.c | 57 +++++++++++++++++++++++++++++++++++ >> 12 files changed, 174 insertions(+), 6 deletions(-) >> >> diff --git a/PROTOCOL b/PROTOCOL >> index ccd3cd2c..e9337e27 100644 >> --- a/PROTOCOL >> +++ b/PROTOCOL >> @@ -427,6 +427,9 @@ Added new command for communication with objects. >> PA_COMMAND_SEND_OBJECT_MESSAGE: >> sends a message to an object that registered a named message handler >> >> +Added signal subscription functions. Clients can now subscribe to signals >> +that PulseAudio sends. > The PROTOCOL file is supposed to document the protocol changes in > detail (or at least that's my impression), but this description is a > very high-level note. Also, talking about "functions" is off-topic > since the protocol doesn't define any functions. > Where does your impression come from? My impression was, that the PROTOCOL file has only very brief and high level descriptions, so I kept it that way. I can surely add more details, if you want.
